In this series of posts, I want to discuss several myths, misconceptions, and misunderstandings that threaten to derail inexperienced or non-technical founders of tech startups.
Note: this post also appears on Glowdot with permission.
In the beginning stages of planning an app development, there are two ubiquitous questions that get asked, and they are the obvious ones:
- How much will it cost to build this app?
- How long will it take to build this app?
These are not unreasonable questions — indeed, when building anything the first two things you need to wrap your head around are budget and timeline.
However, these questions become much more difficult — if not impossible — to answer when it comes to software. Let’s look at why that is.
Software is not physical
To start, when we set out to build software, we really need to understand that we aren’t building a physical, standalone thing that exists in its own space, on its own terms, independent of anything else.
Unlike, say, a house or a car which may need occasional maintenance or upgrades, software often needs constant maintenance and upgrades. It also often needs to evolve. There are a few common reasons for this.
The platform your software runs on has changed
Apple and Google constantly roll out updates to the software (iOS and Android) that powers our devices. These updates can include improvements to the overall performance of our devices, but they can also include breaking changes that cause our apps to suddenly not work as expected. This can include things like changes to way the OS allows apps to interface with hardware (permissions, drivers, etc) or changes to the SDK that we use to perform OS exposed tasks (deprecated methods, etc).
The APIs your software uses have changed
If your app relies on third party APIs — as many do — then you need to respond to changes in those APIs often.
A very common example is the Facebook SDK, which changes often and results in Facebook often retiring old methods. Sometimes these changes can be handled easily by updating the SDK and rewriting a few lines of code. Sometimes these changes can fundamentally change the way your app works.
A few years ago, Facebook changed the way apps could request a user’s friend data. And by “changed the way” I really mean, took away the option. You may remember a time when mobile games were spamming your Facebook wall with requests for you to join so-and-so in a game. When Facebook took away the ability for apps to request a users friends (without special permission to do so), many, many apps had to completely change the way they handled on-boarding and user acquisition.
Users are not static
Your users change, because people change. In order to stay relevant, your software is going to have to evolve to keep up with user expectations.
UI/UX expectations change
In the many years I have been building software, I have seen tremendous changes in the expectations of users in terms of not only functionality and performance, but even more importantly user interface (UI) and user experience (UX).
Its important to remember that the dominant apps — the Facebooks and Instagrams and so on — have whole teams devoted to improving the UI/UX of their apps, and even publishing papers on new theories regarding human computer interaction. When Instagram changes their interface, there is often a shockwave that affects every other app out in the wild, as the UI paradigm shifts in another direction.
User requirements change
Users’ needs change often! Let’s say you have an app that allows people to upload images, get a link, and post the link to their images in various places.
This works great, until a new social network comes along, and they forbid posting links. They do allow posting a special kind of code however. We have seen this in the past in the form of WordPress shortcodes, BBCode, and so on.
If that site blows up, you will definitely want your app to adapt and allow your users to post their images on this new site. If you don’t, I can assure you someone else will come along and fill that user need!
Your expectations change!
There is another even more common change that happens — you change your mind! It happens much, mucb more often than you might think. As development progresses and you get to hold in your hands this thing that started as just an idea, you start get a different perspective on it. You realize there are things you didn’t consider before. You realize some features you thought were critical are actually useless or at least less important. You simply reconsider and adjust, and therefore change direction. Sometimes, you may pivot and completely change the concept entirely!
Although in the best case we always set out to build the thing we originally wanted to build, its good to keep an open mind and remain flexible.
Bug free software is not a thing!
It pains me to say this, but there really is no such thing as bug free software. This is especially true as your software grows in scope and complexity.
What we strive to achieve, really, is a minimum set of bugs. There is not a single app you use that doesn’t have a giant bug list that is waiting to be tackled. If I could pull back the curtain on any piece of software, and it doesn’t matter what type of software, what platform, or who built it, there is a dedicated bug tracking system behind the scenes in which people are constantly entering bugs and developers are constantly pulling out bug reports and trying to fix them. Its just a reality of software development.
That bing said, however, your users should in any case be blissfully unaware of the number of issues in the queue — and often, they are. Many of the bugs you see in these lists are minor issues or issues that affect only a tiny subset of users doing very specific things. However, they still linger and need to be addressed — and as your software evolves, I can almost guarantee new ones will pop up.
Changes trigger new requirements
This is a big one. A change in one place often requires many changes in many other places. Being able to predict those new requirements comes with experience.
For example, lets say the CEO of a blog platform app hears from users that they want to be able to upload photos with their posts. He asks the CTO how long it will take to do it.
The inexperienced CTO thinks about it, realizes adding the client side functionality to choose and upload a photo is pretty simple, and the backend capability of receiving and storing the photo is also pretty simple, and linking the two together by adding a column to the Posts table in the database is a piece of cake, so he answers: “that’ll take a couple days!”.
The experienced CTO realizes that adding photo uploads creates a whole new set of requirements beyond just uploading a photo. We need to be able to delete photos, and replace photos. We need to be able to moderate user uploads to keep unwanted content off our platform — this alone often requires building a whole new software bundle to allow content moderators to do their work! We need to think about things like validating files on the client side and the server side (making sure users don’t upload files that aren’t images). We probably want to add some logic to resize uploads so that if a popular user uploads a 1GB image we don’t have to serve that giant file to everyone who views their post.
And on and on and on. That one simple feature request cascades out into a whole new set of requirements — and possibly those new requirements create their own requirements!
And the list goes on
There are even more reasons we can discuss, but hopefully I’ve made the point.
As a final example, the first big product I ever built was a social network and media sharing site that grew alongside MySpace in the early 2000s.
It started small — just a form to upload photos and get a sharing code running on a single, humble server. Eventually it evolved into a full-fledged social network, and grew to the point where it ran in a stack of highly powerful servers in a dedicated datacenter.
Let’s use that site to answer the original question here: when was development “finished”? When it was retired 12 years after it was launched! New features were constantly being added or removed according to user demands. New enhancements were constantly being made as traffic increased. It truly was a living, evolving being that needed constant care and maintenance, and that constant care and maintenance is largely why it was so successful.
So what’s the answer?
I’ve discussed in part why the original two questions that started this blog post are difficult if not impossible to answer. But I also noted that those questions are crucial to ask before starting any development! So how do we answer those questions?
The key is getting specific. Instead of asking “how much does it cost to build this app?” we break down the lifespan of the product into chunks we can wrap our head around — workshops, prototypes, alpha/beta development phases, release candidates and, after release, versioned updates. The question then becomes, “how long will it take to prototype an app with the following features….” or “how much will it take to get an MVP of my app, on iOS and Android, with this limited feature set“?
In this way, we can more accurately make estimates about how long and at what cost a project will take. Often, the first quote and proposal I offer is a package of several of those questions — we plan out as far ahead as is reasonable. This usually means a few workshops to set the plan, a prototype, an alpha and beta development where we plan out a few critical “must have” features, and then a final pre-release phase where we make sure everything we need to release the app publicly — aka the MVP or minimum viable product — and then we stop there.
Once we reach release, then we can regroup, discuss where we are, put together a list of feature requests, changes, and bug fixes, and estimate a timeline and budget for that new chunk of development. Not only does this help us answer those two hard questions more accurately, it helps us stay focused and on track by encouraging us all to not bite off more than we can chew.
By keeping the size and scope of each phase manageable, we set attainable goals rather than looking too far into the future and risk getting lost along the way.
There is an extra bit of knowledge here: if you are shopping around for quotes and you ask generally “how much will it cost to make this app?” and you get an explicit answer — run and don’t look back. This is a developer that is telling you what they think you want to hear, not what you need to hear. These developers will do anything to get you to sign on with them, from telling you anything and everything is possible, agreeing to every feature request without any advice as to how it might be approached differently, more efficiently, and more cost effectively, and they will approach your project as a single task instead of the complex procedure that it is.
Last modified: February 25, 2019